The current assembly line balancing studies ignore the preventive maintenance (PM) of machines in some workstations, implying that the already-known PM information has been completely missed. Moreover, PM may bring about a production stoppage for a considerable time. Hence, this paper considers PM scenarios into the assembly line balancing problem to improve the production efficiency and smoothness simultaneously. For this multi-objective problem, a heuristic rule relying on the tacit knowledge is dug up via gene expression programming to obtain an acceptable solution quickly. Then, an enhanced grey wolf optimizer algorithm with two improvements is proposed to achieve Pareto front solutions. Specifically, a variable step-size decoding mechanism accelerates the speed of the algorithm; the specially-designed neighbor operators prevent the algorithm from trapping in local optima. Experiment results demonstrate that the discovered heuristic rule outperforms other existing rules; the joint of improvements endows the proposed meta-heuristic with significant superiority over three variants and other six well-known algorithms. Besides, a real-world case study is conducted to validate the discovered rule and the proposed meta-heuristic.

当前的装配线平衡研究忽略了某些工作站中机器的预防性维护（PM），这意味着已经完全错过了已知的PM信息。而且，PM可能导致相当长的时间停产。因此，本文将PM方案考虑到装配线平衡问题中，以同时提高生产效率和平滑度。对于此多目标问题，通过基因表达编程来挖掘依赖于默认信息的启发式规则，以快速获得可接受的解决方案。然后，提出了具有两个改进的增强型灰狼优化器算法，以实现Pareto前沿解。具体而言，可变步长解码机制可加快算法的速度；特别设计的邻居算子可防止算法陷入局部最优状态。实验结果表明，所发现的启发式规则优于其他现有规则。改进的结合赋予了拟议的元启发式方法明显优于三个变体和其他六个著名算法的优势。此外，还进行了一个实际案例研究，以验证发现的规则和提出的元启发式算法。